ERROR ANALYSIS

Chemistry 141 - Fall 2016

Error and uncertainty may seem synonymous with trivial mistakes in the lab, but they are
actually well-defined aspects of any numerical measurement in a laboratory experiment.

A number reported without consideration of precision tells an incomplete story, thus a goal in the
General Chemistry laboratory is to start thinking about the value and precision of any numerical
result—whether it is a laboratory measurement, a survey result, or a sports statistic.

The use of significant digits in calculations is covered in the textbook and is used in the lecture
portion of General Chemistry. Significant digits are an approximate approach to treat precision,
and they provide rules for addition and multiplication.

This guide will be sufficient for your work in General Chemistry. More information can be found
in John R. Taylor’s An Introduction to Error Analysis.*

Random and Systematic Error

One goal for lab work will be controlling the two types of experimental error: systematic
error and random error.

Systematic error arises from a flaw in experimental design or equipment and can be
detected and corrected. This type of error leads to inaccurate measurements of the true value.
The best way to check for systematic error is to use different methods to perform the same
measurement.

Random error is always present and cannot be corrected. It has to do with the precision of
measurements in laboratory and is the statistical uncertainty in the last digits of the precision.
An example of random error is that which arises from reading a burette, which is somewhat
subjective and therefore varies at each reading. Note: “Error” and “uncertainty” are
sometimes used interchangeably to mean “random error.” The phrase “error in a
measurement” is synonymous with “uncertainty in a measurement.”

One aim will be to eliminate systematic error and minimize random error to obtain a high
degree of both accuracy and precision. A goal of the General Chemistry laboratory will be to
practice thinking about the largest contributors to both types of error in our experiments.
 CH141 Laboratory Error Analysis Fall 2016 page 2 of 9

Random Error in Lab Experiments

Systematic error is corrected for in the lab procedure. Understanding random error comes
from repeated measurements to give a set of replicas. The statistics of the set of replicas
gives a way to understand the value and error in measurement. The statistical tools to be used
are the Mean, the Standard Deviation, and the Standard Deviation of the Mean.

Mean, Standard Deviation, and Standard Deviation of the Mean

The mean, , is the simple average of n replicas, each replica signified as .

The standard deviation (SD), , is related to the spread in values of replicas in an

experiment. We will use the population standard deviation to approximate it:

The standard deviation of the mean (SDOM), , is also called the estimated standard
error. If the full experiment was repeated (all replicas), the spread in the average results
should be related to the standard deviation of the mean:

Using Excel

Here the discussion is limited to the mean, the standard deviation, and the standard deviation of
the mean. For example, when five different volumetric flasks were each used to measure 50.0
mL of water at 25 ˚C and then each quantity of water was weighed, the following masses in
grams were determined: 49.897, 49.938, 49.599, 49.965, 49.771. A spreadsheet was used to
calculate the mean, standard deviation and standard deviation of the mean as shown in Figure 1.
 CH141 Laboratory Error Analysis Fall 2016 page 3 of 9

Figure 1. The actual appearance of the spreadsheet used to do the calculations is shown to
the left and the same spreadsheet appears on the right with the equations used to calculate
the quantities shown instead of the numerical result of those calculations.

The number of decimal places shown for the mean, SD, and SDOM were adjusted with the
buttons in the toolbar and using the SDOM to decide the last decimal place to report.
Remember the %error is equal to (SDOM/mean) *100.

Reporting Error
Whenever possible report numerical results with error. The quantity we’ll use to estimate
measurement error in General Chemistry is the SDOM, which will be reported with one
significant digit. This precision of the error will determine the precision of the measurement
(e.g., indicates the last decimal place to the right to report).

Reporting Absolute Error

The absolute error has the same units as the measurement. In the example above, the mass
of 50.0 mL of water at 25 ˚C would be reported as 49.8 ± 0.1 g or 4.98± 0.01´101 g in
scientific notation. The 0.1 g is the error bar on this measurement and signifies the last
decimal place in the measurement to report.

Reporting Relative Error (% error)

The relative error is the scale of the error with respect to the value of the measurement.
Mathematically, it is the SDOM divided by the mean times 100: ( / ) ´ 100%. The
relative error or % error is reported with units of percentage, and it typically has one
significant digit. Thus, for the above example, this quantity was calculated in Excel (see
Figure 1) thus the mass of 50.0 mL of water at 25 ˚C would be reported as 49.8 g ± 0.2%.
 CH141 Laboratory Error Analysis Fall 2016 page 4 of 9

Significance

In General Chemistry lab, the word “significant” doesn’t always mean “important.” The
calculated error bars will be used to gauge the significance of measurements. If two measured
values are contained within each other’s error bars, the measured values are the same—with no
significant difference between them. If two measured values and their error bars are well
separated, then the two values are unique, and their difference is “significant.” There is a grey
area between those two extremes, and some results can be inconclusive. A full understanding of
the significance of inconclusive results requires other statistical tests such as a t-test. More
information can be found on page 150 of Taylor.1

Multiplying and Adding with Error

Once a measurement has been made some dimensional analysis needs to be done. There are easy
rules to follow for multiplication and addition:

Multiplication/Division
Multiplication and division scale a quantity, so these operations also scale the error. The
mass of water in the example above can be used to calculate the density of water at 25 ˚C,
by dividing the mass by the volume. The absolute error will be divided by the volume as
well:
1 g
49.8 ± 0.1 g ✕ = 0.996 ± 0.002
50.0 mL mL

Addition/Subtraction
Addition and subtraction just shift a quantity, so they don’t change the error. For instance,
if the temperature of the room was measured to be 25.3 ± 0.3 ˚C, and that value was then
converted to Kelvin by adding 273.15 to the measured temperature, as shown below. The
absolute error is unchanged because Kelvin and Celsius have the same scale, just a
different origin.
25.3 ± 0.3 ˚C + 273.15 = 298.4 ± 0.3 K
 CH141 Laboratory Error Analysis Fall 2016 page 5 of 9

Error in Linear Fits with Excel

One last place to get errors from measurements is from the linear fits done by Excel. For
example, using volumetric glassware, different volumes of water were measured and weighed at
25 ˚C. The following graph was made following the instructions in the Excel Help document.

Mass versus Volume of Water at 25 ˚C

!
Figure 2. The mass of water at 25 ˚C increases linearly with volume, giving a slope of
0.994 ±0.002 g/mL and a y-intercept of 0.06 ± 0.08 g.

Where did ±0.002 g/mL and ± 0.08 g in the figure caption come from? While it’s good to give
the R2-value to give a sense for the precision of the fit a better strategy is to use Excel’s LINEST
function to calculate the errors in these parameters.

To Use Excel’s LINEST Function

1. Highlight an empty block of cells two columns wide (↔) by five rows (↕) high.
2. Immediately type =LINEST(
3. Highlight the cells containing the y-values and then type a comma ,
4. Highlight the cells containing the x-values and then type a comma ,
5. Type one more comma ,
6. Type TRUE) so your cell should look similar to: =LINEST(B3:B7,A3:A7„TRUE)
 CH141 Laboratory Error Analysis Fall 2016 page 6 of 9

7. Hit the command and return buttons

8. The Excel table below shows exactly what the LINEST generates for output. The top two
rows of the LINEST function output contain the value and error for the slope and intercept.
These data were needed to make the Figure 2 caption above. Please make sure that the r2
reported for the slope is consistent with the value obtained from the equation for the line in
your graph. If you wish, you can look up the meanings of the other quantities reported by
LINEST as they are beyond the scope of this course and document.

Figure 3. LINEST output for mass of water versus volume correlation.

Reference
*Taylor, J. R. An Introduction to to Error Analysis: The Study of Uncertainties in Physical
Measurements, 2nd Ed.; University Science Books: Sausalito, CA, 1997.
 CH141 Laboratory Error Analysis Fall 2016 page 7 of 9

Error Analysis Examples

Error analysis is often a difficult area for students. However, the careful consideration of
experimental error is one of the important skills that we need to learn to be effective scientists. In
the following discussion, the errors in a titration experiment are considered. The first section is a
detailed look at how to determine the most important errors. The second section is an example of
the corresponding text that would be written in a lab report for CH141-142.
Determining the Important Errors
• The purpose of the error analysis section of the lab report is to determine the most important
errors and the effect that those errors have on the final result.
• Random Errors: Random errors cause positive and negative deviations from the average
value of a measurement. Random errors cancel by averaging, if the experiment is repeated many
times. Upon averaging many trials, random errors have an effect only on the precision of a
measurement. The effect of random errors is primarily on the precision. Every non-integer
experimental measurement is a source of random error. The random error is estimated from the
readability of the device. A table of typical measurements and the associated precision, under
practical circumstances, is given below. For instrument readings, to avoid round-off error, report
one extra significant figure and then underline the digit that is not significant.

Volumetric flasks and pipettes precision (relative) significant figures

10 mL ± 0.03 mL 3 i.e. 10.0 mL
25 mL ± 0.03 mL 3 25.0 mL
50 mL ± 0.05 mL 3 50.0 mL
100 mL ± 0.08 mL 4 100.0 mL
Auto-pipettors
10 µL ± 0.05 µL (0.5%) 3 i.e. 10.0 µL
100 µL ± 0.3 µL (0.3%) 3 100. µL
1000 µL ± 2 µL (0.2%) 3 1000. µL
Analytical Balance
0.1000 g ± 0.0001 g 4 e.g. 0.3456 g
1.0000 g ± 0.0001 g 5 2.3456 g
10.0000 g ± 0.0001 g 6 12.3456 g
Spectrophotometer (Spectro-Viz*)
0.100 ± 0.003 (± 3%) 2 e.g. 0.123
0.500 ± 0.003 (± 0.6%) 2 0.456
1.000 ± 0.004 (± 0.4%) 3 1.123
1.500 (1 sign. fig. for A>2) ± 0.015 (± 1%) 2 1.567
pH Meter
7.00 ± 0.02 ~3 e.g. 6.123
Electronic Pressure Sensor
1.0 atm (760 torr) ± 0.0005 atm (± 0.4 torr) 3-4 e.g 826.4 torr
Constant Current Power Supply
0.400 amp ± 0.0004 amp (± 0.1 %) 3 0.4162 amp
* Spectro-Viz plus photometric accuracy is ± 13%, but standard curve calibration decreases the systematic error to
approximately equal the averaged precision (about 3 sign. figures), assuming the range of A is 0.1 to 1.0.
 CH141 Laboratory Error Analysis Fall 2016 page 8 of 9

• Systematic Errors: Without any changes in the procedure, systematic errors are repeated if
the experiment is repeated. Systematic errors have a biased effect on the final results; systematic
errors make the final result high or low, but not both. Instrument calibration errors are examples
of systematic errors. Environmental effects can also be causes of systematic error, for example a
change in lab temperature changing the calibration of a balance or the volume of a flask. An
example of a systematic error from the CaCO3 precipitation experiment is that small particles
pass through the glass frits in a Gooch crucible, making the final precipitate mass too small.
Systematic errors affect the accuracy of the final results.
A given measurement can contribute to both random and systematic error. Non-integer
measurements always contribute to the random error. For example, a miscalibrated balance is a
source of random and systematic error. Systematic errors are often corrected by completing a
determination using a different method or by comparing results among different laboratories.
• Student Mistakes: Student mistakes are just student mistakes; they are neither random nor
systematic errors. Examples in this category are spills, misreading a device such as a burette,
misinterpretation of the procedure, incorrect handling of a micro-pipettor, and forgetting to rinse
out a beaker when doing a quantitative transfer. These errors are known and easily preventable,
if the experiment is repeated. Systematic errors occur with each repetition of the experiment,
assuming no changes in instrumentation. Mistakes should be noted in the Results section of your
report as mistakes.

Example: Titration of an Unknown Acid:

A 25.0-mL sample of an unknown acid is titrated with 15.67 mL of 0.1042 M NaOH. The
volume of the acid is determined using a volumetric pipette and the burette used in the
experiment has scale divisions every 0.1 mL. The standard base solution was made using an
analytical balance and a 100.0-mL volumetric flask. The end point is determined by visually
detecting the pink color of phenolphthalein.
Answer:
Random Measurement Errors: Every measurement is a source of random error. However, we
must identify those errors that have a significant effect on the final result. The effects on the final
result are determined using significant figure rules. The concentration of the unknown acid is:
Munknown = VtitrantMtitrant/Vunknown = 0.01567 L(0.1042 mol/L)/0.0250 L = 0.4821 M
Since only multiplications and divisions are involved, the number of significant figures in the
final result is equal to the smallest number of significant figures of the terms in the calculation.
We next discuss the errors associated with each term.
Volumetric Glassware and Analytical Balance Measurements: Large volume standard
volumetric glassware typically has a precision and accuracy of four significant figures. The
accuracy and precision of mass measurements on an analytical balance are also typically to four
significant figures (±0.0001). The expected precision in the NaOH solution, using the analytical
balance and volumetric glassware, is four significant figures. The number of significant figures
in the 25.0-mL volumetric pipette is three. At best the final concentration is known to three
significant figures.
Measurements that are Interpolated between Scale Markings: The burette readings are not as
precise. To determine the volume of titrant delivered, two readings are made. Each reading is
 CH141 Laboratory Error Analysis Fall 2016 page 9 of 9

recorded to the nearest 0.01 mL. However, visually estimating the volume to better than ±0.02
mL is difficult. Consequently the precision of the volume delivered by the burette is poorer than
±0.02 mL, since two readings are necessary. Correspondingly in the titration example, the
volume delivered by the buret at best is 15.67 ± 0.03 mL, or three significant figures.
Correspondingly, the final unknown concentration is officially known to three significant
figures. The conclusion is that the precision is determined primarily by the random error in the
burette readings and pipette. The random error in the other volume and mass determinations are
not consequential.
Systematic Measurement Errors: Every measurement is a potential source of systematic error.
However, with thoughtful construction of the procedure many measurements can be discounted
as significant sources of systematic error. So while the calibration of the glassware and the
balance used in a titration experiment are technically sources of systematic error, these errors are
easily avoided. The calibration of the balances is periodically checked using a registered
calibration mass. Standard volumetric glassware is certified by the manufacturer through
calibration against National Institute of Standards and Technology (NIST) traceable procedures.
In a titration experiment the only significant systematic errors are in the purity of the reagents
and the visual determination of the end point. The purity of the reagents also includes absorption
of moisture from the ambient air. So reagents that are susceptible to atmospheric moisture
absorption are usually kept in low humidity desiccators. In a titration, the primary systematic
error is the endpoint determination. The difference between the equivalence point and the
measured end point is called the titration error. A visual end point is always slightly beyond the
equivalence point because of the necessity of seeing the color change by eye. The result is that
the volume of titrant delivered is too large, giving a larger final concentration than the true value.
The conclusion is that the accuracy is determined primarily by systematic error in the end point.
Example Lab Report Section on Error Analysis
The discussion, above, gives the complete thought process for determining the most important
errors in the experiment. The section in the lab report that presents your conclusions is
disappointingly short, by comparison. For the titration example:

The precision is dominated by the random error of the volume readings of the burette and
volumetric pipette. The other volumetric glassware contribute insignificant random error.
The accuracy is determined by the systematic error in the visual detection of the end point.
The visual end point is at a volume larger than the equivalence point, giving a higher final
result than the true concentration value.